| /* Target-dependent code for SPARC. |
| |
| Copyright (C) 2003-2024 Free Software Foundation, Inc. |
| |
| This file is part of GDB. |
| |
| This program is free software; you can redistribute it and/or modify |
| it under the terms of the GNU General Public License as published by |
| the Free Software Foundation; either version 3 of the License, or |
| (at your option) any later version. |
| |
| This program is distributed in the hope that it will be useful, |
| but WITHOUT ANY WARRANTY; without even the implied warranty of |
| MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| GNU General Public License for more details. |
| |
| You should have received a copy of the GNU General Public License |
| along with this program. If not, see <http://www.gnu.org/licenses/>. */ |
| |
| #include "arch-utils.h" |
| #include "dis-asm.h" |
| #include "dwarf2.h" |
| #include "dwarf2/frame.h" |
| #include "extract-store-integer.h" |
| #include "frame.h" |
| #include "frame-base.h" |
| #include "frame-unwind.h" |
| #include "gdbcore.h" |
| #include "gdbtypes.h" |
| #include "inferior.h" |
| #include "symtab.h" |
| #include "objfiles.h" |
| #include "osabi.h" |
| #include "regcache.h" |
| #include "target.h" |
| #include "target-descriptions.h" |
| #include "value.h" |
| |
| #include "sparc-tdep.h" |
| #include "sparc-ravenscar-thread.h" |
| #include <algorithm> |
| |
| struct regset; |
| |
| /* This file implements the SPARC 32-bit ABI as defined by the section |
| "Low-Level System Information" of the SPARC Compliance Definition |
| (SCD) 2.4.1, which is the 32-bit System V psABI for SPARC. The SCD |
| lists changes with respect to the original 32-bit psABI as defined |
| in the "System V ABI, SPARC Processor Supplement". |
| |
| Note that if we talk about SunOS, we mean SunOS 4.x, which was |
| BSD-based, which is sometimes (retroactively?) referred to as |
| Solaris 1.x. If we talk about Solaris we mean Solaris 2.x and |
| above (Solaris 7, 8 and 9 are nothing but Solaris 2.7, 2.8 and 2.9 |
| suffering from severe version number inflation). Solaris 2.x is |
| also known as SunOS 5.x, since that's what uname(1) says. Solaris |
| 2.x is SVR4-based. */ |
| |
| /* Please use the sparc32_-prefix for 32-bit specific code, the |
| sparc64_-prefix for 64-bit specific code and the sparc_-prefix for |
| code that can handle both. The 64-bit specific code lives in |
| sparc64-tdep.c; don't add any here. */ |
| |
| /* The stack pointer is offset from the stack frame by a BIAS of 2047 |
| (0x7ff) for 64-bit code. BIAS is likely to be defined on SPARC |
| hosts, so undefine it first. */ |
| #undef BIAS |
| #define BIAS 2047 |
| |
| /* Macros to extract fields from SPARC instructions. */ |
| #define X_OP(i) (((i) >> 30) & 0x3) |
| #define X_RD(i) (((i) >> 25) & 0x1f) |
| #define X_A(i) (((i) >> 29) & 1) |
| #define X_COND(i) (((i) >> 25) & 0xf) |
| #define X_OP2(i) (((i) >> 22) & 0x7) |
| #define X_IMM22(i) ((i) & 0x3fffff) |
| #define X_OP3(i) (((i) >> 19) & 0x3f) |
| #define X_RS1(i) (((i) >> 14) & 0x1f) |
| #define X_RS2(i) ((i) & 0x1f) |
| #define X_I(i) (((i) >> 13) & 1) |
| /* Sign extension macros. */ |
| #define X_DISP22(i) ((X_IMM22 (i) ^ 0x200000) - 0x200000) |
| #define X_DISP19(i) ((((i) & 0x7ffff) ^ 0x40000) - 0x40000) |
| #define X_DISP10(i) ((((((i) >> 11) && 0x300) | (((i) >> 5) & 0xff)) ^ 0x200) - 0x200) |
| #define X_SIMM13(i) ((((i) & 0x1fff) ^ 0x1000) - 0x1000) |
| /* Macros to identify some instructions. */ |
| /* RETURN (RETT in V8) */ |
| #define X_RETTURN(i) ((X_OP (i) == 0x2) && (X_OP3 (i) == 0x39)) |
| |
| /* Fetch the instruction at PC. Instructions are always big-endian |
| even if the processor operates in little-endian mode. */ |
| |
| unsigned long |
| sparc_fetch_instruction (CORE_ADDR pc) |
| { |
| gdb_byte buf[4]; |
| unsigned long insn; |
| int i; |
| |
| /* If we can't read the instruction at PC, return zero. */ |
| if (target_read_memory (pc, buf, sizeof (buf))) |
| return 0; |
| |
| insn = 0; |
| for (i = 0; i < sizeof (buf); i++) |
| insn = (insn << 8) | buf[i]; |
| return insn; |
| } |
| |
| |
| /* Return non-zero if the instruction corresponding to PC is an "unimp" |
| instruction. */ |
| |
| static int |
| sparc_is_unimp_insn (CORE_ADDR pc) |
| { |
| const unsigned long insn = sparc_fetch_instruction (pc); |
| |
| return ((insn & 0xc1c00000) == 0); |
| } |
| |
| /* Return non-zero if the instruction corresponding to PC is an |
| "annulled" branch, i.e. the annul bit is set. */ |
| |
| int |
| sparc_is_annulled_branch_insn (CORE_ADDR pc) |
| { |
| /* The branch instructions featuring an annul bit can be identified |
| by the following bit patterns: |
| |
| OP=0 |
| OP2=1: Branch on Integer Condition Codes with Prediction (BPcc). |
| OP2=2: Branch on Integer Condition Codes (Bcc). |
| OP2=5: Branch on FP Condition Codes with Prediction (FBfcc). |
| OP2=6: Branch on FP Condition Codes (FBcc). |
| OP2=3 && Bit28=0: |
| Branch on Integer Register with Prediction (BPr). |
| |
| This leaves out ILLTRAP (OP2=0), SETHI/NOP (OP2=4) and the V8 |
| coprocessor branch instructions (Op2=7). */ |
| |
| const unsigned long insn = sparc_fetch_instruction (pc); |
| const unsigned op2 = X_OP2 (insn); |
| |
| if ((X_OP (insn) == 0) |
| && ((op2 == 1) || (op2 == 2) || (op2 == 5) || (op2 == 6) |
| || ((op2 == 3) && ((insn & 0x10000000) == 0)))) |
| return X_A (insn); |
| else |
| return 0; |
| } |
| |
| /* OpenBSD/sparc includes StackGhost, which according to the author's |
| website http://stackghost.cerias.purdue.edu "... transparently and |
| automatically protects applications' stack frames; more |
| specifically, it guards the return pointers. The protection |
| mechanisms require no application source or binary modification and |
| imposes only a negligible performance penalty." |
| |
| The same website provides the following description of how |
| StackGhost works: |
| |
| "StackGhost interfaces with the kernel trap handler that would |
| normally write out registers to the stack and the handler that |
| would read them back in. By XORing a cookie into the |
| return-address saved in the user stack when it is actually written |
| to the stack, and then XOR it out when the return-address is pulled |
| from the stack, StackGhost can cause attacker corrupted return |
| pointers to behave in a manner the attacker cannot predict. |
| StackGhost can also use several unused bits in the return pointer |
| to detect a smashed return pointer and abort the process." |
| |
| For GDB this means that whenever we're reading %i7 from a stack |
| frame's window save area, we'll have to XOR the cookie. |
| |
| More information on StackGuard can be found on in: |
| |
| Mike Frantzen and Mike Shuey. "StackGhost: Hardware Facilitated |
| Stack Protection." 2001. Published in USENIX Security Symposium |
| '01. */ |
| |
| /* Fetch StackGhost Per-Process XOR cookie. */ |
| |
| ULONGEST |
| sparc_fetch_wcookie (struct gdbarch *gdbarch) |
| { |
| enum bfd_endian byte_order = gdbarch_byte_order (gdbarch); |
| struct target_ops *ops = current_inferior ()->top_target (); |
| gdb_byte buf[8]; |
| int len; |
| |
| len = target_read (ops, TARGET_OBJECT_WCOOKIE, NULL, buf, 0, 8); |
| if (len == -1) |
| return 0; |
| |
| /* We should have either an 32-bit or an 64-bit cookie. */ |
| gdb_assert (len == 4 || len == 8); |
| |
| return extract_unsigned_integer (buf, len, byte_order); |
| } |
| |
| |
| /* The functions on this page are intended to be used to classify |
| function arguments. */ |
| |
| /* Check whether TYPE is "Integral or Pointer". */ |
| |
| static int |
| sparc_integral_or_pointer_p (const struct type *type) |
| { |
| int len = type->length (); |
| |
| switch (type->code ()) |
| { |
| case TYPE_CODE_INT: |
| case TYPE_CODE_BOOL: |
| case TYPE_CODE_CHAR: |
| case TYPE_CODE_ENUM: |
| case TYPE_CODE_RANGE: |
| /* We have byte, half-word, word and extended-word/doubleword |
| integral types. The doubleword is an extension to the |
| original 32-bit ABI by the SCD 2.4.x. */ |
| return (len == 1 || len == 2 || len == 4 || len == 8); |
| case TYPE_CODE_PTR: |
| case TYPE_CODE_REF: |
| case TYPE_CODE_RVALUE_REF: |
| /* Allow either 32-bit or 64-bit pointers. */ |
| return (len == 4 || len == 8); |
| default: |
| break; |
| } |
| |
| return 0; |
| } |
| |
| /* Check whether TYPE is "Floating". */ |
| |
| static int |
| sparc_floating_p (const struct type *type) |
| { |
| switch (type->code ()) |
| { |
| case TYPE_CODE_FLT: |
| { |
| int len = type->length (); |
| return (len == 4 || len == 8 || len == 16); |
| } |
| default: |
| break; |
| } |
| |
| return 0; |
| } |
| |
| /* Check whether TYPE is "Complex Floating". */ |
| |
| static int |
| sparc_complex_floating_p (const struct type *type) |
| { |
| switch (type->code ()) |
| { |
| case TYPE_CODE_COMPLEX: |
| { |
| int len = type->length (); |
| return (len == 8 || len == 16 || len == 32); |
| } |
| default: |
| break; |
| } |
| |
| return 0; |
| } |
| |
| /* Check whether TYPE is "Structure or Union". |
| |
| In terms of Ada subprogram calls, arrays are treated the same as |
| struct and union types. So this function also returns non-zero |
| for array types. */ |
| |
| static int |
| sparc_structure_or_union_p (const struct type *type) |
| { |
| switch (type->code ()) |
| { |
| case TYPE_CODE_STRUCT: |
| case TYPE_CODE_UNION: |
| case TYPE_CODE_ARRAY: |
| return 1; |
| default: |
| break; |
| } |
| |
| return 0; |
| } |
| |
| /* Return true if TYPE is returned by memory, false if returned by |
| register. */ |
| |
| static bool |
| sparc_structure_return_p (const struct type *type) |
| { |
| if (type->code () == TYPE_CODE_ARRAY && type->is_vector ()) |
| { |
| /* Float vectors are always returned by memory. */ |
| if (sparc_floating_p (check_typedef (type->target_type ()))) |
| return true; |
| /* Integer vectors are returned by memory if the vector size |
| is greater than 8 bytes long. */ |
| return (type->length () > 8); |
| } |
| |
| if (sparc_floating_p (type)) |
| { |
| /* Floating point types are passed by register for size 4 and |
| 8 bytes, and by memory for size 16 bytes. */ |
| return (type->length () == 16); |
| } |
| |
| /* Other than that, only aggregates of all sizes get returned by |
| memory. */ |
| return sparc_structure_or_union_p (type); |
| } |
| |
| /* Return true if arguments of the given TYPE are passed by |
| memory; false if returned by register. */ |
| |
| static bool |
| sparc_arg_by_memory_p (const struct type *type) |
| { |
| if (type->code () == TYPE_CODE_ARRAY && type->is_vector ()) |
| { |
| /* Float vectors are always passed by memory. */ |
| if (sparc_floating_p (check_typedef (type->target_type ()))) |
| return true; |
| /* Integer vectors are passed by memory if the vector size |
| is greater than 8 bytes long. */ |
| return (type->length () > 8); |
| } |
| |
| /* Floats are passed by register for size 4 and 8 bytes, and by memory |
| for size 16 bytes. */ |
| if (sparc_floating_p (type)) |
| return (type->length () == 16); |
| |
| /* Complex floats and aggregates of all sizes are passed by memory. */ |
| if (sparc_complex_floating_p (type) || sparc_structure_or_union_p (type)) |
| return true; |
| |
| /* Everything else gets passed by register. */ |
| return false; |
| } |
| |
| /* Register information. */ |
| #define SPARC32_FPU_REGISTERS \ |
| "f0", "f1", "f2", "f3", "f4", "f5", "f6", "f7", \ |
| "f8", "f9", "f10", "f11", "f12", "f13", "f14", "f15", \ |
| "f16", "f17", "f18", "f19", "f20", "f21", "f22", "f23", \ |
| "f24", "f25", "f26", "f27", "f28", "f29", "f30", "f31" |
| #define SPARC32_CP0_REGISTERS \ |
| "y", "psr", "wim", "tbr", "pc", "npc", "fsr", "csr" |
| |
| static const char * const sparc_core_register_names[] = { |
| SPARC_CORE_REGISTERS |
| }; |
| static const char * const sparc32_fpu_register_names[] = { |
| SPARC32_FPU_REGISTERS |
| }; |
| static const char * const sparc32_cp0_register_names[] = { |
| SPARC32_CP0_REGISTERS |
| }; |
| |
| static const char * const sparc32_register_names[] = |
| { |
| SPARC_CORE_REGISTERS, |
| SPARC32_FPU_REGISTERS, |
| SPARC32_CP0_REGISTERS |
| }; |
| |
| /* Total number of registers. */ |
| #define SPARC32_NUM_REGS ARRAY_SIZE (sparc32_register_names) |
| |
| /* We provide the aliases %d0..%d30 for the floating registers as |
| "psuedo" registers. */ |
| |
| static const char * const sparc32_pseudo_register_names[] = |
| { |
| "d0", "d2", "d4", "d6", "d8", "d10", "d12", "d14", |
| "d16", "d18", "d20", "d22", "d24", "d26", "d28", "d30" |
| }; |
| |
| /* Total number of pseudo registers. */ |
| #define SPARC32_NUM_PSEUDO_REGS ARRAY_SIZE (sparc32_pseudo_register_names) |
| |
| /* Return the name of pseudo register REGNUM. */ |
| |
| static const char * |
| sparc32_pseudo_register_name (struct gdbarch *gdbarch, int regnum) |
| { |
| regnum -= gdbarch_num_regs (gdbarch); |
| |
| gdb_assert (regnum < SPARC32_NUM_PSEUDO_REGS); |
| return sparc32_pseudo_register_names[regnum]; |
| } |
| |
| /* Return the name of register REGNUM. */ |
| |
| static const char * |
| sparc32_register_name (struct gdbarch *gdbarch, int regnum) |
| { |
| if (tdesc_has_registers (gdbarch_target_desc (gdbarch))) |
| return tdesc_register_name (gdbarch, regnum); |
| |
| if (regnum >= 0 && regnum < gdbarch_num_regs (gdbarch)) |
| return sparc32_register_names[regnum]; |
| |
| return sparc32_pseudo_register_name (gdbarch, regnum); |
| } |
| |
| /* Construct types for ISA-specific registers. */ |
| |
| static struct type * |
| sparc_psr_type (struct gdbarch *gdbarch) |
| { |
| sparc_gdbarch_tdep *tdep = gdbarch_tdep<sparc_gdbarch_tdep> (gdbarch); |
| |
| if (!tdep->sparc_psr_type) |
| { |
| struct type *type; |
| |
| type = arch_flags_type (gdbarch, "builtin_type_sparc_psr", 32); |
| append_flags_type_flag (type, 5, "ET"); |
| append_flags_type_flag (type, 6, "PS"); |
| append_flags_type_flag (type, 7, "S"); |
| append_flags_type_flag (type, 12, "EF"); |
| append_flags_type_flag (type, 13, "EC"); |
| |
| tdep->sparc_psr_type = type; |
| } |
| |
| return tdep->sparc_psr_type; |
| } |
| |
| static struct type * |
| sparc_fsr_type (struct gdbarch *gdbarch) |
| { |
| sparc_gdbarch_tdep *tdep = gdbarch_tdep<sparc_gdbarch_tdep> (gdbarch); |
| |
| if (!tdep->sparc_fsr_type) |
| { |
| struct type *type; |
| |
| type = arch_flags_type (gdbarch, "builtin_type_sparc_fsr", 32); |
| append_flags_type_flag (type, 0, "NXA"); |
| append_flags_type_flag (type, 1, "DZA"); |
| append_flags_type_flag (type, 2, "UFA"); |
| append_flags_type_flag (type, 3, "OFA"); |
| append_flags_type_flag (type, 4, "NVA"); |
| append_flags_type_flag (type, 5, "NXC"); |
| append_flags_type_flag (type, 6, "DZC"); |
| append_flags_type_flag (type, 7, "UFC"); |
| append_flags_type_flag (type, 8, "OFC"); |
| append_flags_type_flag (type, 9, "NVC"); |
| append_flags_type_flag (type, 22, "NS"); |
| append_flags_type_flag (type, 23, "NXM"); |
| append_flags_type_flag (type, 24, "DZM"); |
| append_flags_type_flag (type, 25, "UFM"); |
| append_flags_type_flag (type, 26, "OFM"); |
| append_flags_type_flag (type, 27, "NVM"); |
| |
| tdep->sparc_fsr_type = type; |
| } |
| |
| return tdep->sparc_fsr_type; |
| } |
| |
| /* Return the GDB type object for the "standard" data type of data in |
| pseudo register REGNUM. */ |
| |
| static struct type * |
| sparc32_pseudo_register_type (struct gdbarch *gdbarch, int regnum) |
| { |
| regnum -= gdbarch_num_regs (gdbarch); |
| |
| if (regnum >= SPARC32_D0_REGNUM && regnum <= SPARC32_D30_REGNUM) |
| return builtin_type (gdbarch)->builtin_double; |
| |
| internal_error (_("sparc32_pseudo_register_type: bad register number %d"), |
| regnum); |
| } |
| |
| /* Return the GDB type object for the "standard" data type of data in |
| register REGNUM. */ |
| |
| static struct type * |
| sparc32_register_type (struct gdbarch *gdbarch, int regnum) |
| { |
| if (tdesc_has_registers (gdbarch_target_desc (gdbarch))) |
| return tdesc_register_type (gdbarch, regnum); |
| |
| if (regnum >= SPARC_F0_REGNUM && regnum <= SPARC_F31_REGNUM) |
| return builtin_type (gdbarch)->builtin_float; |
| |
| if (regnum == SPARC_SP_REGNUM || regnum == SPARC_FP_REGNUM) |
| return builtin_type (gdbarch)->builtin_data_ptr; |
| |
| if (regnum == SPARC32_PC_REGNUM || regnum == SPARC32_NPC_REGNUM) |
| return builtin_type (gdbarch)->builtin_func_ptr; |
| |
| if (regnum == SPARC32_PSR_REGNUM) |
| return sparc_psr_type (gdbarch); |
| |
| if (regnum == SPARC32_FSR_REGNUM) |
| return sparc_fsr_type (gdbarch); |
| |
| if (regnum >= gdbarch_num_regs (gdbarch)) |
| return sparc32_pseudo_register_type (gdbarch, regnum); |
| |
| return builtin_type (gdbarch)->builtin_int32; |
| } |
| |
| static enum register_status |
| sparc32_pseudo_register_read (struct gdbarch *gdbarch, |
| readable_regcache *regcache, |
| int regnum, gdb_byte *buf) |
| { |
| enum register_status status; |
| |
| regnum -= gdbarch_num_regs (gdbarch); |
| gdb_assert (regnum >= SPARC32_D0_REGNUM && regnum <= SPARC32_D30_REGNUM); |
| |
| regnum = SPARC_F0_REGNUM + 2 * (regnum - SPARC32_D0_REGNUM); |
| status = regcache->raw_read (regnum, buf); |
| if (status == REG_VALID) |
| status = regcache->raw_read (regnum + 1, buf + 4); |
| return status; |
| } |
| |
| static void |
| sparc32_pseudo_register_write (struct gdbarch *gdbarch, |
| struct regcache *regcache, |
| int regnum, const gdb_byte *buf) |
| { |
| regnum -= gdbarch_num_regs (gdbarch); |
| gdb_assert (regnum >= SPARC32_D0_REGNUM && regnum <= SPARC32_D30_REGNUM); |
| |
| regnum = SPARC_F0_REGNUM + 2 * (regnum - SPARC32_D0_REGNUM); |
| regcache->raw_write (regnum, buf); |
| regcache->raw_write (regnum + 1, buf + 4); |
| } |
| |
| /* Implement the stack_frame_destroyed_p gdbarch method. */ |
| |
| int |
| sparc_stack_frame_destroyed_p (struct gdbarch *gdbarch, CORE_ADDR pc) |
| { |
| /* This function must return true if we are one instruction after an |
| instruction that destroyed the stack frame of the current |
| function. The SPARC instructions used to restore the callers |
| stack frame are RESTORE and RETURN/RETT. |
| |
| Of these RETURN/RETT is a branch instruction and thus we return |
| true if we are in its delay slot. |
| |
| RESTORE is almost always found in the delay slot of a branch |
| instruction that transfers control to the caller, such as JMPL. |
| Thus the next instruction is in the caller frame and we don't |
| need to do anything about it. */ |
| |
| unsigned int insn = sparc_fetch_instruction (pc - 4); |
| |
| return X_RETTURN (insn); |
| } |
| |
| |
| static CORE_ADDR |
| sparc32_frame_align (struct gdbarch *gdbarch, CORE_ADDR address) |
| { |
| /* The ABI requires double-word alignment. */ |
| return address & ~0x7; |
| } |
| |
| static CORE_ADDR |
| sparc32_push_dummy_code (struct gdbarch *gdbarch, CORE_ADDR sp, |
| CORE_ADDR funcaddr, |
| struct value **args, int nargs, |
| struct type *value_type, |
| CORE_ADDR *real_pc, CORE_ADDR *bp_addr, |
| struct regcache *regcache) |
| { |
| enum bfd_endian byte_order = gdbarch_byte_order (gdbarch); |
| |
| *bp_addr = sp - 4; |
| *real_pc = funcaddr; |
| |
| if (using_struct_return (gdbarch, NULL, value_type)) |
| { |
| gdb_byte buf[4]; |
| |
| /* This is an UNIMP instruction. */ |
| store_unsigned_integer (buf, 4, byte_order, |
| value_type->length () & 0x1fff); |
| write_memory (sp - 8, buf, 4); |
| return sp - 8; |
| } |
| |
| return sp - 4; |
| } |
| |
| static CORE_ADDR |
| sparc32_store_arguments (struct regcache *regcache, int nargs, |
| struct value **args, CORE_ADDR sp, |
| function_call_return_method return_method, |
| CORE_ADDR struct_addr) |
| { |
| struct gdbarch *gdbarch = regcache->arch (); |
| enum bfd_endian byte_order = gdbarch_byte_order (gdbarch); |
| /* Number of words in the "parameter array". */ |
| int num_elements = 0; |
| int element = 0; |
| int i; |
| |
| for (i = 0; i < nargs; i++) |
| { |
| struct type *type = args[i]->type (); |
| int len = type->length (); |
| |
| if (sparc_arg_by_memory_p (type)) |
| { |
| /* Structure, Union and Quad-Precision Arguments. */ |
| sp -= len; |
| |
| /* Use doubleword alignment for these values. That's always |
| correct, and wasting a few bytes shouldn't be a problem. */ |
| sp &= ~0x7; |
| |
| write_memory (sp, args[i]->contents ().data (), len); |
| args[i] = value_from_pointer (lookup_pointer_type (type), sp); |
| num_elements++; |
| } |
| else if (sparc_floating_p (type)) |
| { |
| /* Floating arguments. */ |
| gdb_assert (len == 4 || len == 8); |
| num_elements += (len / 4); |
| } |
| else |
| { |
| /* Arguments passed via the General Purpose Registers. */ |
| num_elements += ((len + 3) / 4); |
| } |
| } |
| |
| /* Always allocate at least six words. */ |
| sp -= std::max (6, num_elements) * 4; |
| |
| /* The psABI says that "Software convention requires space for the |
| struct/union return value pointer, even if the word is unused." */ |
| sp -= 4; |
| |
| /* The psABI says that "Although software convention and the |
| operating system require every stack frame to be doubleword |
| aligned." */ |
| sp &= ~0x7; |
| |
| for (i = 0; i < nargs; i++) |
| { |
| const bfd_byte *valbuf = args[i]->contents ().data (); |
| struct type *type = args[i]->type (); |
| int len = type->length (); |
| gdb_byte buf[4]; |
| |
| if (len < 4) |
| { |
| memset (buf, 0, 4 - len); |
| memcpy (buf + 4 - len, valbuf, len); |
| valbuf = buf; |
| len = 4; |
| } |
| |
| gdb_assert (len == 4 || len == 8); |
| |
| if (element < 6) |
| { |
| int regnum = SPARC_O0_REGNUM + element; |
| |
| regcache->cooked_write (regnum, valbuf); |
| if (len > 4 && element < 5) |
| regcache->cooked_write (regnum + 1, valbuf + 4); |
| } |
| |
| /* Always store the argument in memory. */ |
| write_memory (sp + 4 + element * 4, valbuf, len); |
| element += len / 4; |
| } |
| |
| gdb_assert (element == num_elements); |
| |
| if (return_method == return_method_struct) |
| { |
| gdb_byte buf[4]; |
| |
| store_unsigned_integer (buf, 4, byte_order, struct_addr); |
| write_memory (sp, buf, 4); |
| } |
| |
| return sp; |
| } |
| |
| static CORE_ADDR |
| sparc32_push_dummy_call (struct gdbarch *gdbarch, struct value *function, |
| struct regcache *regcache, CORE_ADDR bp_addr, |
| int nargs, struct value **args, CORE_ADDR sp, |
| function_call_return_method return_method, |
| CORE_ADDR struct_addr) |
| { |
| CORE_ADDR call_pc = (return_method == return_method_struct |
| ? (bp_addr - 12) : (bp_addr - 8)); |
| |
| /* Set return address. */ |
| regcache_cooked_write_unsigned (regcache, SPARC_O7_REGNUM, call_pc); |
| |
| /* Set up function arguments. */ |
| sp = sparc32_store_arguments (regcache, nargs, args, sp, return_method, |
| struct_addr); |
| |
| /* Allocate the 16-word window save area. */ |
| sp -= 16 * 4; |
| |
| /* Stack should be doubleword aligned at this point. */ |
| gdb_assert (sp % 8 == 0); |
| |
| /* Finally, update the stack pointer. */ |
| regcache_cooked_write_unsigned (regcache, SPARC_SP_REGNUM, sp); |
| |
| return sp; |
| } |
| |
| |
| /* Use the program counter to determine the contents and size of a |
| breakpoint instruction. Return a pointer to a string of bytes that |
| encode a breakpoint instruction, store the length of the string in |
| *LEN and optionally adjust *PC to point to the correct memory |
| location for inserting the breakpoint. */ |
| constexpr gdb_byte sparc_break_insn[] = { 0x91, 0xd0, 0x20, 0x01 }; |
| |
| typedef BP_MANIPULATION (sparc_break_insn) sparc_breakpoint; |
| |
| |
| /* Allocate and initialize a frame cache. */ |
| |
| static struct sparc_frame_cache * |
| sparc_alloc_frame_cache (void) |
| { |
| struct sparc_frame_cache *cache; |
| |
| cache = FRAME_OBSTACK_ZALLOC (struct sparc_frame_cache); |
| |
| /* Base address. */ |
| cache->base = 0; |
| cache->pc = 0; |
| |
| /* Frameless until proven otherwise. */ |
| cache->frameless_p = 1; |
| cache->frame_offset = 0; |
| cache->saved_regs_mask = 0; |
| cache->copied_regs_mask = 0; |
| cache->struct_return_p = 0; |
| |
| return cache; |
| } |
| |
| /* GCC generates several well-known sequences of instructions at the begining |
| of each function prologue when compiling with -fstack-check. If one of |
| such sequences starts at START_PC, then return the address of the |
| instruction immediately past this sequence. Otherwise, return START_PC. */ |
| |
| static CORE_ADDR |
| sparc_skip_stack_check (const CORE_ADDR start_pc) |
| { |
| CORE_ADDR pc = start_pc; |
| unsigned long insn; |
| int probing_loop = 0; |
| |
| /* With GCC, all stack checking sequences begin with the same two |
| instructions, plus an optional one in the case of a probing loop: |
| |
| sethi <some immediate>, %g1 |
| sub %sp, %g1, %g1 |
| |
| or: |
| |
| sethi <some immediate>, %g1 |
| sethi <some immediate>, %g4 |
| sub %sp, %g1, %g1 |
| |
| or: |
| |
| sethi <some immediate>, %g1 |
| sub %sp, %g1, %g1 |
| sethi <some immediate>, %g4 |
| |
| If the optional instruction is found (setting g4), assume that a |
| probing loop will follow. */ |
| |
| /* sethi <some immediate>, %g1 */ |
| insn = sparc_fetch_instruction (pc); |
| pc = pc + 4; |
| if (!(X_OP (insn) == 0 && X_OP2 (insn) == 0x4 && X_RD (insn) == 1)) |
| return start_pc; |
| |
| /* optional: sethi <some immediate>, %g4 */ |
| insn = sparc_fetch_instruction (pc); |
| pc = pc + 4; |
| if (X_OP (insn) == 0 && X_OP2 (insn) == 0x4 && X_RD (insn) == 4) |
| { |
| probing_loop = 1; |
| insn = sparc_fetch_instruction (pc); |
| pc = pc + 4; |
| } |
| |
| /* sub %sp, %g1, %g1 */ |
| if (!(X_OP (insn) == 2 && X_OP3 (insn) == 0x4 && !X_I(insn) |
| && X_RD (insn) == 1 && X_RS1 (insn) == 14 && X_RS2 (insn) == 1)) |
| return start_pc; |
| |
| insn = sparc_fetch_instruction (pc); |
| pc = pc + 4; |
| |
| /* optional: sethi <some immediate>, %g4 */ |
| if (X_OP (insn) == 0 && X_OP2 (insn) == 0x4 && X_RD (insn) == 4) |
| { |
| probing_loop = 1; |
| insn = sparc_fetch_instruction (pc); |
| pc = pc + 4; |
| } |
| |
| /* First possible sequence: |
| [first two instructions above] |
| clr [%g1 - some immediate] */ |
| |
| /* clr [%g1 - some immediate] */ |
| if (X_OP (insn) == 3 && X_OP3(insn) == 0x4 && X_I(insn) |
| && X_RS1 (insn) == 1 && X_RD (insn) == 0) |
| { |
| /* Valid stack-check sequence, return the new PC. */ |
| return pc; |
| } |
| |
| /* Second possible sequence: A small number of probes. |
| [first two instructions above] |
| clr [%g1] |
| add %g1, -<some immediate>, %g1 |
| clr [%g1] |
| [repeat the two instructions above any (small) number of times] |
| clr [%g1 - some immediate] */ |
| |
| /* clr [%g1] */ |
| else if (X_OP (insn) == 3 && X_OP3(insn) == 0x4 && !X_I(insn) |
| && X_RS1 (insn) == 1 && X_RD (insn) == 0) |
| { |
| while (1) |
| { |
| /* add %g1, -<some immediate>, %g1 */ |
| insn = sparc_fetch_instruction (pc); |
| pc = pc + 4; |
| if (!(X_OP (insn) == 2 && X_OP3(insn) == 0 && X_I(insn) |
| && X_RS1 (insn) == 1 && X_RD (insn) == 1)) |
| break; |
| |
| /* clr [%g1] */ |
| insn = sparc_fetch_instruction (pc); |
| pc = pc + 4; |
| if (!(X_OP (insn) == 3 && X_OP3(insn) == 0x4 && !X_I(insn) |
| && X_RD (insn) == 0 && X_RS1 (insn) == 1)) |
| return start_pc; |
| } |
| |
| /* clr [%g1 - some immediate] */ |
| if (!(X_OP (insn) == 3 && X_OP3(insn) == 0x4 && X_I(insn) |
| && X_RS1 (insn) == 1 && X_RD (insn) == 0)) |
| return start_pc; |
| |
| /* We found a valid stack-check sequence, return the new PC. */ |
| return pc; |
| } |
| |
| /* Third sequence: A probing loop. |
| [first three instructions above] |
| sub %g1, %g4, %g4 |
| cmp %g1, %g4 |
| be <disp> |
| add %g1, -<some immediate>, %g1 |
| ba <disp> |
| clr [%g1] |
| |
| And an optional last probe for the remainder: |
| |
| clr [%g4 - some immediate] */ |
| |
| if (probing_loop) |
| { |
| /* sub %g1, %g4, %g4 */ |
| if (!(X_OP (insn) == 2 && X_OP3 (insn) == 0x4 && !X_I(insn) |
| && X_RD (insn) == 4 && X_RS1 (insn) == 1 && X_RS2 (insn) == 4)) |
| return start_pc; |
| |
| /* cmp %g1, %g4 */ |
| insn = sparc_fetch_instruction (pc); |
| pc = pc + 4; |
| if (!(X_OP (insn) == 2 && X_OP3 (insn) == 0x14 && !X_I(insn) |
| && X_RD (insn) == 0 && X_RS1 (insn) == 1 && X_RS2 (insn) == 4)) |
| return start_pc; |
| |
| /* be <disp> */ |
| insn = sparc_fetch_instruction (pc); |
| pc = pc + 4; |
| if (!(X_OP (insn) == 0 && X_COND (insn) == 0x1)) |
| return start_pc; |
| |
| /* add %g1, -<some immediate>, %g1 */ |
| insn = sparc_fetch_instruction (pc); |
| pc = pc + 4; |
| if (!(X_OP (insn) == 2 && X_OP3(insn) == 0 && X_I(insn) |
| && X_RS1 (insn) == 1 && X_RD (insn) == 1)) |
| return start_pc; |
| |
| /* ba <disp> */ |
| insn = sparc_fetch_instruction (pc); |
| pc = pc + 4; |
| if (!(X_OP (insn) == 0 && X_COND (insn) == 0x8)) |
| return start_pc; |
| |
| /* clr [%g1] (st %g0, [%g1] or st %g0, [%g1+0]) */ |
| insn = sparc_fetch_instruction (pc); |
| pc = pc + 4; |
| if (!(X_OP (insn) == 3 && X_OP3(insn) == 0x4 |
| && X_RD (insn) == 0 && X_RS1 (insn) == 1 |
| && (!X_I(insn) || X_SIMM13 (insn) == 0))) |
| return start_pc; |
| |
| /* We found a valid stack-check sequence, return the new PC. */ |
| |
| /* optional: clr [%g4 - some immediate] */ |
| insn = sparc_fetch_instruction (pc); |
| pc = pc + 4; |
| if (!(X_OP (insn) == 3 && X_OP3(insn) == 0x4 && X_I(insn) |
| && X_RS1 (insn) == 4 && X_RD (insn) == 0)) |
| return pc - 4; |
| else |
| return pc; |
| } |
| |
| /* No stack check code in our prologue, return the start_pc. */ |
| return start_pc; |
| } |
| |
| /* Record the effect of a SAVE instruction on CACHE. */ |
| |
| void |
| sparc_record_save_insn (struct sparc_frame_cache *cache) |
| { |
| /* The frame is set up. */ |
| cache->frameless_p = 0; |
| |
| /* The frame pointer contains the CFA. */ |
| cache->frame_offset = 0; |
| |
| /* The `local' and `in' registers are all saved. */ |
| cache->saved_regs_mask = 0xffff; |
| |
| /* The `out' registers are all renamed. */ |
| cache->copied_regs_mask = 0xff; |
| } |
| |
| /* Do a full analysis of the prologue at PC and update CACHE accordingly. |
| Bail out early if CURRENT_PC is reached. Return the address where |
| the analysis stopped. |
| |
| We handle both the traditional register window model and the single |
| register window (aka flat) model. */ |
| |
| CORE_ADDR |
| sparc_analyze_prologue (struct gdbarch *gdbarch, CORE_ADDR pc, |
| CORE_ADDR current_pc, struct sparc_frame_cache *cache) |
| { |
| sparc_gdbarch_tdep *tdep = gdbarch_tdep<sparc_gdbarch_tdep> (gdbarch); |
| unsigned long insn; |
| int offset = 0; |
| int dest = -1; |
| |
| pc = sparc_skip_stack_check (pc); |
| |
| if (current_pc <= pc) |
| return current_pc; |
| |
| /* We have to handle to "Procedure Linkage Table" (PLT) special. On |
| SPARC the linker usually defines a symbol (typically |
| _PROCEDURE_LINKAGE_TABLE_) at the start of the .plt section. |
| This symbol makes us end up here with PC pointing at the start of |
| the PLT and CURRENT_PC probably pointing at a PLT entry. If we |
| would do our normal prologue analysis, we would probably conclude |
| that we've got a frame when in reality we don't, since the |
| dynamic linker patches up the first PLT with some code that |
| starts with a SAVE instruction. Patch up PC such that it points |
| at the start of our PLT entry. */ |
| if (tdep->plt_entry_size > 0 && in_plt_section (current_pc)) |
| pc = current_pc - ((current_pc - pc) % tdep->plt_entry_size); |
| |
| insn = sparc_fetch_instruction (pc); |
| |
| /* Recognize store insns and record their sources. */ |
| while (X_OP (insn) == 3 |
| && (X_OP3 (insn) == 0x4 /* stw */ |
| || X_OP3 (insn) == 0x7 /* std */ |
| || X_OP3 (insn) == 0xe) /* stx */ |
| && X_RS1 (insn) == SPARC_SP_REGNUM) |
| { |
| int regnum = X_RD (insn); |
| |
| /* Recognize stores into the corresponding stack slots. */ |
| if (regnum >= SPARC_L0_REGNUM && regnum <= SPARC_I7_REGNUM |
| && ((X_I (insn) |
| && X_SIMM13 (insn) == (X_OP3 (insn) == 0xe |
| ? (regnum - SPARC_L0_REGNUM) * 8 + BIAS |
| : (regnum - SPARC_L0_REGNUM) * 4)) |
| || (!X_I (insn) && regnum == SPARC_L0_REGNUM))) |
| { |
| cache->saved_regs_mask |= (1 << (regnum - SPARC_L0_REGNUM)); |
| if (X_OP3 (insn) == 0x7) |
| cache->saved_regs_mask |= (1 << (regnum + 1 - SPARC_L0_REGNUM)); |
| } |
| |
| offset += 4; |
| |
| insn = sparc_fetch_instruction (pc + offset); |
| } |
| |
| /* Recognize a SETHI insn and record its destination. */ |
| if (X_OP (insn) == 0 && X_OP2 (insn) == 0x04) |
| { |
| dest = X_RD (insn); |
| offset += 4; |
| |
| insn = sparc_fetch_instruction (pc + offset); |
| } |
| |
| /* Allow for an arithmetic operation on DEST or %g1. */ |
| if (X_OP (insn) == 2 && X_I (insn) |
| && (X_RD (insn) == 1 || X_RD (insn) == dest)) |
| { |
| offset += 4; |
| |
| insn = sparc_fetch_instruction (pc + offset); |
| } |
| |
| /* Check for the SAVE instruction that sets up the frame. */ |
| if (X_OP (insn) == 2 && X_OP3 (insn) == 0x3c) |
| { |
| sparc_record_save_insn (cache); |
| offset += 4; |
| return pc + offset; |
| } |
| |
| /* Check for an arithmetic operation on %sp. */ |
| if (X_OP (insn) == 2 |
| && (X_OP3 (insn) == 0 || X_OP3 (insn) == 0x4) |
| && X_RS1 (insn) == SPARC_SP_REGNUM |
| && X_RD (insn) == SPARC_SP_REGNUM) |
| { |
| if (X_I (insn)) |
| { |
| cache->frame_offset = X_SIMM13 (insn); |
| if (X_OP3 (insn) == 0) |
| cache->frame_offset = -cache->frame_offset; |
| } |
| offset += 4; |
| |
| insn = sparc_fetch_instruction (pc + offset); |
| |
| /* Check for an arithmetic operation that sets up the frame. */ |
| if (X_OP (insn) == 2 |
| && (X_OP3 (insn) == 0 || X_OP3 (insn) == 0x4) |
| && X_RS1 (insn) == SPARC_SP_REGNUM |
| && X_RD (insn) == SPARC_FP_REGNUM) |
| { |
| cache->frameless_p = 0; |
| cache->frame_offset = 0; |
| /* We could check that the amount subtracted to %sp above is the |
| same as the one added here, but this seems superfluous. */ |
| cache->copied_regs_mask |= 0x40; |
| offset += 4; |
| |
| insn = sparc_fetch_instruction (pc + offset); |
| } |
| |
| /* Check for a move (or) operation that copies the return register. */ |
| if (X_OP (insn) == 2 |
| && X_OP3 (insn) == 0x2 |
| && !X_I (insn) |
| && X_RS1 (insn) == SPARC_G0_REGNUM |
| && X_RS2 (insn) == SPARC_O7_REGNUM |
| && X_RD (insn) == SPARC_I7_REGNUM) |
| { |
| cache->copied_regs_mask |= 0x80; |
| offset += 4; |
| } |
| |
| return pc + offset; |
| } |
| |
| return pc; |
| } |
| |
| /* Return PC of first real instruction of the function starting at |
| START_PC. */ |
| |
| static CORE_ADDR |
| sparc32_skip_prologue (struct gdbarch *gdbarch, CORE_ADDR start_pc) |
| { |
| CORE_ADDR func_addr; |
| struct sparc_frame_cache cache; |
| |
| /* This is the preferred method, find the end of the prologue by |
| using the debugging information. */ |
| |
| if (find_pc_partial_function (start_pc, NULL, &func_addr, NULL)) |
| { |
| CORE_ADDR post_prologue_pc |
| = skip_prologue_using_sal (gdbarch, func_addr); |
| |
| if (post_prologue_pc != 0) |
| return std::max (start_pc, post_prologue_pc); |
| } |
| |
| start_pc = sparc_analyze_prologue (gdbarch, start_pc, 0xffffffffUL, &cache); |
| |
| /* The psABI says that "Although the first 6 words of arguments |
| reside in registers, the standard stack frame reserves space for |
| them.". It also suggests that a function may use that space to |
| "write incoming arguments 0 to 5" into that space, and that's |
| indeed what GCC seems to be doing. In that case GCC will |
| generate debug information that points to the stack slots instead |
| of the registers, so we should consider the instructions that |
| write out these incoming arguments onto the stack. */ |
| |
| while (1) |
| { |
| unsigned long insn = sparc_fetch_instruction (start_pc); |
| |
| /* Recognize instructions that store incoming arguments into the |
| corresponding stack slots. */ |
| if (X_OP (insn) == 3 && (X_OP3 (insn) & 0x3c) == 0x04 |
| && X_I (insn) && X_RS1 (insn) == SPARC_FP_REGNUM) |
| { |
| int regnum = X_RD (insn); |
| |
| /* Case of arguments still in %o[0..5]. */ |
| if (regnum >= SPARC_O0_REGNUM && regnum <= SPARC_O5_REGNUM |
| && !(cache.copied_regs_mask & (1 << (regnum - SPARC_O0_REGNUM))) |
| && X_SIMM13 (insn) == 68 + (regnum - SPARC_O0_REGNUM) * 4) |
| { |
| start_pc += 4; |
| continue; |
| } |
| |
| /* Case of arguments copied into %i[0..5]. */ |
| if (regnum >= SPARC_I0_REGNUM && regnum <= SPARC_I5_REGNUM |
| && (cache.copied_regs_mask & (1 << (regnum - SPARC_I0_REGNUM))) |
| && X_SIMM13 (insn) == 68 + (regnum - SPARC_I0_REGNUM) * 4) |
| { |
| start_pc += 4; |
| continue; |
| } |
| } |
| |
| break; |
| } |
| |
| return start_pc; |
| } |
| |
| /* Normal frames. */ |
| |
| struct sparc_frame_cache * |
| sparc_frame_cache (const frame_info_ptr &this_frame, void **this_cache) |
| { |
| struct sparc_frame_cache *cache; |
| |
| if (*this_cache) |
| return (struct sparc_frame_cache *) *this_cache; |
| |
| cache = sparc_alloc_frame_cache (); |
| *this_cache = cache; |
| |
| cache->pc = get_frame_func (this_frame); |
| if (cache->pc != 0) |
| sparc_analyze_prologue (get_frame_arch (this_frame), cache->pc, |
| get_frame_pc (this_frame), cache); |
| |
| if (cache->frameless_p) |
| { |
| /* This function is frameless, so %fp (%i6) holds the frame |
| pointer for our calling frame. Use %sp (%o6) as this frame's |
| base address. */ |
| cache->base = |
| get_frame_register_unsigned (this_frame, SPARC_SP_REGNUM); |
| } |
| else |
| { |
| /* For normal frames, %fp (%i6) holds the frame pointer, the |
| base address for the current stack frame. */ |
| cache->base = |
| get_frame_register_unsigned (this_frame, SPARC_FP_REGNUM); |
| } |
| |
| cache->base += cache->frame_offset; |
| |
| if (cache->base & 1) |
| cache->base += BIAS; |
| |
| return cache; |
| } |
| |
| static int |
| sparc32_struct_return_from_sym (struct symbol *sym) |
| { |
| struct type *type = check_typedef (sym->type ()); |
| enum type_code code = type->code (); |
| |
| if (code == TYPE_CODE_FUNC || code == TYPE_CODE_METHOD) |
| { |
| type = check_typedef (type->target_type ()); |
| if (sparc_structure_or_union_p (type) |
| || (sparc_floating_p (type) && type->length () == 16)) |
| return 1; |
| } |
| |
| return 0; |
| } |
| |
| struct sparc_frame_cache * |
| sparc32_frame_cache (const frame_info_ptr &this_frame, void **this_cache) |
| { |
| struct sparc_frame_cache *cache; |
| struct symbol *sym; |
| |
| if (*this_cache) |
| return (struct sparc_frame_cache *) *this_cache; |
| |
| cache = sparc_frame_cache (this_frame, this_cache); |
| |
| sym = find_pc_function (cache->pc); |
| if (sym) |
| { |
| cache->struct_return_p = sparc32_struct_return_from_sym (sym); |
| } |
| else |
| { |
| /* There is no debugging information for this function to |
| help us determine whether this function returns a struct |
| or not. So we rely on another heuristic which is to check |
| the instruction at the return address and see if this is |
| an "unimp" instruction. If it is, then it is a struct-return |
| function. */ |
| CORE_ADDR pc; |
| int regnum = |
| (cache->copied_regs_mask & 0x80) ? SPARC_I7_REGNUM : SPARC_O7_REGNUM; |
| |
| pc = get_frame_register_unsigned (this_frame, regnum) + 8; |
| if (sparc_is_unimp_insn (pc)) |
| cache->struct_return_p = 1; |
| } |
| |
| return cache; |
| } |
| |
| static void |
| sparc32_frame_this_id (const frame_info_ptr &this_frame, void **this_cache, |
| struct frame_id *this_id) |
| { |
| struct sparc_frame_cache *cache = |
| sparc32_frame_cache (this_frame, this_cache); |
| |
| /* This marks the outermost frame. */ |
| if (cache->base == 0) |
| return; |
| |
| (*this_id) = frame_id_build (cache->base, cache->pc); |
| } |
| |
| static struct value * |
| sparc32_frame_prev_register (const frame_info_ptr &this_frame, |
| void **this_cache, int regnum) |
| { |
| struct gdbarch *gdbarch = get_frame_arch (this_frame); |
| struct sparc_frame_cache *cache = |
| sparc32_frame_cache (this_frame, this_cache); |
| |
| if (regnum == SPARC32_PC_REGNUM || regnum == SPARC32_NPC_REGNUM) |
| { |
| CORE_ADDR pc = (regnum == SPARC32_NPC_REGNUM) ? 4 : 0; |
| |
| /* If this functions has a Structure, Union or Quad-Precision |
| return value, we have to skip the UNIMP instruction that encodes |
| the size of the structure. */ |
| if (cache->struct_return_p) |
| pc += 4; |
| |
| regnum = |
| (cache->copied_regs_mask & 0x80) ? SPARC_I7_REGNUM : SPARC_O7_REGNUM; |
| pc += get_frame_register_unsigned (this_frame, regnum) + 8; |
| return frame_unwind_got_constant (this_frame, regnum, pc); |
| } |
| |
| /* Handle StackGhost. */ |
| { |
| ULONGEST wcookie = sparc_fetch_wcookie (gdbarch); |
| |
| if (wcookie != 0 && !cache->frameless_p && regnum == SPARC_I7_REGNUM) |
| { |
| CORE_ADDR addr = cache->base + (regnum - SPARC_L0_REGNUM) * 4; |
| ULONGEST i7; |
| |
| /* Read the value in from memory. */ |
| i7 = get_frame_memory_unsigned (this_frame, addr, 4); |
| return frame_unwind_got_constant (this_frame, regnum, i7 ^ wcookie); |
| } |
| } |
| |
| /* The previous frame's `local' and `in' registers may have been saved |
| in the register save area. */ |
| if (regnum >= SPARC_L0_REGNUM && regnum <= SPARC_I7_REGNUM |
| && (cache->saved_regs_mask & (1 << (regnum - SPARC_L0_REGNUM)))) |
| { |
| CORE_ADDR addr = cache->base + (regnum - SPARC_L0_REGNUM) * 4; |
| |
| return frame_unwind_got_memory (this_frame, regnum, addr); |
| } |
| |
| /* The previous frame's `out' registers may be accessible as the current |
| frame's `in' registers. */ |
| if (regnum >= SPARC_O0_REGNUM && regnum <= SPARC_O7_REGNUM |
| && (cache->copied_regs_mask & (1 << (regnum - SPARC_O0_REGNUM)))) |
| regnum += (SPARC_I0_REGNUM - SPARC_O0_REGNUM); |
| |
| return frame_unwind_got_register (this_frame, regnum, regnum); |
| } |
| |
| static const struct frame_unwind sparc32_frame_unwind = |
| { |
| "sparc32 prologue", |
| NORMAL_FRAME, |
| default_frame_unwind_stop_reason, |
| sparc32_frame_this_id, |
| sparc32_frame_prev_register, |
| NULL, |
| default_frame_sniffer |
| }; |
| |
| |
| static CORE_ADDR |
| sparc32_frame_base_address (const frame_info_ptr &this_frame, void **this_cache) |
| { |
| struct sparc_frame_cache *cache = |
| sparc32_frame_cache (this_frame, this_cache); |
| |
| return cache->base; |
| } |
| |
| static const struct frame_base sparc32_frame_base = |
| { |
| &sparc32_frame_unwind, |
| sparc32_frame_base_address, |
| sparc32_frame_base_address, |
| sparc32_frame_base_address |
| }; |
| |
| static struct frame_id |
| sparc_dummy_id (struct gdbarch *gdbarch, const frame_info_ptr &this_frame) |
| { |
| CORE_ADDR sp; |
| |
| sp = get_frame_register_unsigned (this_frame, SPARC_SP_REGNUM); |
| if (sp & 1) |
| sp += BIAS; |
| return frame_id_build (sp, get_frame_pc (this_frame)); |
| } |
| |
| |
| /* Extract a function return value of TYPE from REGCACHE, and copy |
| that into VALBUF. */ |
| |
| static void |
| sparc32_extract_return_value (struct type *type, struct regcache *regcache, |
| gdb_byte *valbuf) |
| { |
| int len = type->length (); |
| gdb_byte buf[32]; |
| |
| gdb_assert (!sparc_structure_return_p (type)); |
| |
| if (sparc_floating_p (type) || sparc_complex_floating_p (type) |
| || type->code () == TYPE_CODE_ARRAY) |
| { |
| /* Floating return values. */ |
| regcache->cooked_read (SPARC_F0_REGNUM, buf); |
| if (len > 4) |
| regcache->cooked_read (SPARC_F1_REGNUM, buf + 4); |
| if (len > 8) |
| { |
| regcache->cooked_read (SPARC_F2_REGNUM, buf + 8); |
| regcache->cooked_read (SPARC_F3_REGNUM, buf + 12); |
| } |
| if (len > 16) |
| { |
| regcache->cooked_read (SPARC_F4_REGNUM, buf + 16); |
| regcache->cooked_read (SPARC_F5_REGNUM, buf + 20); |
| regcache->cooked_read (SPARC_F6_REGNUM, buf + 24); |
| regcache->cooked_read (SPARC_F7_REGNUM, buf + 28); |
| } |
| memcpy (valbuf, buf, len); |
| } |
| else |
| { |
| /* Integral and pointer return values. */ |
| gdb_assert (sparc_integral_or_pointer_p (type)); |
| |
| regcache->cooked_read (SPARC_O0_REGNUM, buf); |
| if (len > 4) |
| { |
| regcache->cooked_read (SPARC_O1_REGNUM, buf + 4); |
| gdb_assert (len == 8); |
| memcpy (valbuf, buf, 8); |
| } |
| else |
| { |
| /* Just stripping off any unused bytes should preserve the |
| signed-ness just fine. */ |
| memcpy (valbuf, buf + 4 - len, len); |
| } |
| } |
| } |
| |
| /* Store the function return value of type TYPE from VALBUF into |
| REGCACHE. */ |
| |
| static void |
| sparc32_store_return_value (struct type *type, struct regcache *regcache, |
| const gdb_byte *valbuf) |
| { |
| int len = type->length (); |
| gdb_byte buf[32]; |
| |
| gdb_assert (!sparc_structure_return_p (type)); |
| |
| if (sparc_floating_p (type) || sparc_complex_floating_p (type)) |
| { |
| /* Floating return values. */ |
| memcpy (buf, valbuf, len); |
| regcache->cooked_write (SPARC_F0_REGNUM, buf); |
| if (len > 4) |
| regcache->cooked_write (SPARC_F1_REGNUM, buf + 4); |
| if (len > 8) |
| { |
| regcache->cooked_write (SPARC_F2_REGNUM, buf + 8); |
| regcache->cooked_write (SPARC_F3_REGNUM, buf + 12); |
| } |
| if (len > 16) |
| { |
| regcache->cooked_write (SPARC_F4_REGNUM, buf + 16); |
| regcache->cooked_write (SPARC_F5_REGNUM, buf + 20); |
| regcache->cooked_write (SPARC_F6_REGNUM, buf + 24); |
| regcache->cooked_write (SPARC_F7_REGNUM, buf + 28); |
| } |
| } |
| else |
| { |
| /* Integral and pointer return values. */ |
| gdb_assert (sparc_integral_or_pointer_p (type)); |
| |
| if (len > 4) |
| { |
| gdb_assert (len == 8); |
| memcpy (buf, valbuf, 8); |
| regcache->cooked_write (SPARC_O1_REGNUM, buf + 4); |
| } |
| else |
| { |
| /* ??? Do we need to do any sign-extension here? */ |
| memcpy (buf + 4 - len, valbuf, len); |
| } |
| regcache->cooked_write (SPARC_O0_REGNUM, buf); |
| } |
| } |
| |
| static enum return_value_convention |
| sparc32_return_value (struct gdbarch *gdbarch, struct value *function, |
| struct type *type, struct regcache *regcache, |
| struct value **read_value, const gdb_byte *writebuf) |
| { |
| enum bfd_endian byte_order = gdbarch_byte_order (gdbarch); |
| |
| /* The psABI says that "...every stack frame reserves the word at |
| %fp+64. If a function returns a structure, union, or |
| quad-precision value, this word should hold the address of the |
| object into which the return value should be copied." This |
| guarantees that we can always find the return value, not just |
| before the function returns. */ |
| |
| if (sparc_structure_return_p (type)) |
| { |
| ULONGEST sp; |
| CORE_ADDR addr; |
| |
| if (read_value != nullptr) |
| { |
| regcache_cooked_read_unsigned (regcache, SPARC_SP_REGNUM, &sp); |
| addr = read_memory_unsigned_integer (sp + 64, 4, byte_order); |
| *read_value = value_at_non_lval (type, addr); |
| } |
| if (writebuf) |
| { |
| regcache_cooked_read_unsigned (regcache, SPARC_SP_REGNUM, &sp); |
| addr = read_memory_unsigned_integer (sp + 64, 4, byte_order); |
| write_memory (addr, writebuf, type->length ()); |
| } |
| |
| return RETURN_VALUE_ABI_PRESERVES_ADDRESS; |
| } |
| |
| if (read_value != nullptr) |
| { |
| *read_value = value::allocate (type); |
| gdb_byte *readbuf = (*read_value)->contents_raw ().data (); |
| sparc32_extract_return_value (type, regcache, readbuf); |
| } |
| if (writebuf) |
| sparc32_store_return_value (type, regcache, writebuf); |
| |
| return RETURN_VALUE_REGISTER_CONVENTION; |
| } |
| |
| static int |
| sparc32_stabs_argument_has_addr (struct gdbarch *gdbarch, struct type *type) |
| { |
| return (sparc_structure_or_union_p (type) |
| || (sparc_floating_p (type) && type->length () == 16) |
| || sparc_complex_floating_p (type)); |
| } |
| |
| static int |
| sparc32_dwarf2_struct_return_p (const frame_info_ptr &this_frame) |
| { |
| CORE_ADDR pc = get_frame_address_in_block (this_frame); |
| struct symbol *sym = find_pc_function (pc); |
| |
| if (sym) |
| return sparc32_struct_return_from_sym (sym); |
| return 0; |
| } |
| |
| static void |
| sparc32_dwarf2_frame_init_reg (struct gdbarch *gdbarch, int regnum, |
| struct dwarf2_frame_state_reg *reg, |
| const frame_info_ptr &this_frame) |
| { |
| int off; |
| |
| switch (regnum) |
| { |
| case SPARC_G0_REGNUM: |
| /* Since %g0 is always zero, there is no point in saving it, and |
| people will be inclined omit it from the CFI. Make sure we |
| don't warn about that. */ |
| reg->how = DWARF2_FRAME_REG_SAME_VALUE; |
| break; |
| case SPARC_SP_REGNUM: |
| reg->how = DWARF2_FRAME_REG_CFA; |
| break; |
| case SPARC32_PC_REGNUM: |
| case SPARC32_NPC_REGNUM: |
| reg->how = DWARF2_FRAME_REG_RA_OFFSET; |
| off = 8; |
| if (sparc32_dwarf2_struct_return_p (this_frame)) |
| off += 4; |
| if (regnum == SPARC32_NPC_REGNUM) |
| off += 4; |
| reg->loc.offset = off; |
| break; |
| } |
| } |
| |
| /* Implement the execute_dwarf_cfa_vendor_op method. */ |
| |
| static bool |
| sparc_execute_dwarf_cfa_vendor_op (struct gdbarch *gdbarch, gdb_byte op, |
| struct dwarf2_frame_state *fs) |
| { |
| /* Only DW_CFA_GNU_window_save is expected on SPARC. */ |
| if (op != DW_CFA_GNU_window_save) |
| return false; |
| |
| uint64_t reg; |
| int size = register_size (gdbarch, 0); |
| |
| fs->regs.alloc_regs (32); |
| for (reg = 8; reg < 16; reg++) |
| { |
| fs->regs.reg[reg].how = DWARF2_FRAME_REG_SAVED_REG; |
| fs->regs.reg[reg].loc.reg = reg + 16; |
| } |
| for (reg = 16; reg < 32; reg++) |
| { |
| fs->regs.reg[reg].how = DWARF2_FRAME_REG_SAVED_OFFSET; |
| fs->regs.reg[reg].loc.offset = (reg - 16) * size; |
| } |
| |
| return true; |
| } |
| |
| |
| /* The SPARC Architecture doesn't have hardware single-step support, |
| and most operating systems don't implement it either, so we provide |
| software single-step mechanism. */ |
| |
| static CORE_ADDR |
| sparc_analyze_control_transfer (struct regcache *regcache, |
| CORE_ADDR pc, CORE_ADDR *npc) |
| { |
| unsigned long insn = sparc_fetch_instruction (pc); |
| int conditional_p = X_COND (insn) & 0x7; |
| int branch_p = 0, fused_p = 0; |
| long offset = 0; /* Must be signed for sign-extend. */ |
| |
| if (X_OP (insn) == 0 && X_OP2 (insn) == 3) |
| { |
| if ((insn & 0x10000000) == 0) |
| { |
| /* Branch on Integer Register with Prediction (BPr). */ |
| branch_p = 1; |
| conditional_p = 1; |
| } |
| else |
| { |
| /* Compare and Branch */ |
| branch_p = 1; |
| fused_p = 1; |
| offset = 4 * X_DISP10 (insn); |
| } |
| } |
| else if (X_OP (insn) == 0 && X_OP2 (insn) == 6) |
| { |
| /* Branch on Floating-Point Condition Codes (FBfcc). */ |
| branch_p = 1; |
| offset = 4 * X_DISP22 (insn); |
| } |
| else if (X_OP (insn) == 0 && X_OP2 (insn) == 5) |
| { |
| /* Branch on Floating-Point Condition Codes with Prediction |
| (FBPfcc). */ |
| branch_p = 1; |
| offset = 4 * X_DISP19 (insn); |
| } |
| else if (X_OP (insn) == 0 && X_OP2 (insn) == 2) |
| { |
| /* Branch on Integer Condition Codes (Bicc). */ |
| branch_p = 1; |
| offset = 4 * X_DISP22 (insn); |
| } |
| else if (X_OP (insn) == 0 && X_OP2 (insn) == 1) |
| { |
| /* Branch on Integer Condition Codes with Prediction (BPcc). */ |
| branch_p = 1; |
| offset = 4 * X_DISP19 (insn); |
| } |
| else if (X_OP (insn) == 2 && X_OP3 (insn) == 0x3a) |
| { |
| frame_info_ptr frame = get_current_frame (); |
| |
| /* Trap instruction (TRAP). */ |
| gdbarch *arch = regcache->arch (); |
| sparc_gdbarch_tdep *tdep = gdbarch_tdep<sparc_gdbarch_tdep> (arch); |
| return tdep->step_trap (frame, insn); |
| } |
| |
| /* FIXME: Handle DONE and RETRY instructions. */ |
| |
| if (branch_p) |
| { |
| if (fused_p) |
| { |
| /* Fused compare-and-branch instructions are non-delayed, |
| and do not have an annulling capability. So we need to |
| always set a breakpoint on both the NPC and the branch |
| target address. */ |
| gdb_assert (offset != 0); |
| return pc + offset; |
| } |
| else if (conditional_p) |
| { |
| /* For conditional branches, return nPC + 4 iff the annul |
| bit is 1. */ |
| return (X_A (insn) ? *npc + 4 : 0); |
| } |
| else |
| { |
| /* For unconditional branches, return the target if its |
| specified condition is "always" and return nPC + 4 if the |
| condition is "never". If the annul bit is 1, set *NPC to |
| zero. */ |
| if (X_COND (insn) == 0x0) |
| pc = *npc, offset = 4; |
| if (X_A (insn)) |
| *npc = 0; |
| |
| return pc + offset; |
| } |
| } |
| |
| return 0; |
| } |
| |
| static CORE_ADDR |
| sparc_step_trap (const frame_info_ptr &frame, unsigned long insn) |
| { |
| return 0; |
| } |
| |
| static std::vector<CORE_ADDR> |
| sparc_software_single_step (struct regcache *regcache) |
| { |
| struct gdbarch *arch = regcache->arch (); |
| sparc_gdbarch_tdep *tdep = gdbarch_tdep<sparc_gdbarch_tdep> (arch); |
| CORE_ADDR npc, nnpc; |
| |
| CORE_ADDR pc, orig_npc; |
| std::vector<CORE_ADDR> next_pcs; |
| |
| pc = regcache_raw_get_unsigned (regcache, tdep->pc_regnum); |
| orig_npc = npc = regcache_raw_get_unsigned (regcache, tdep->npc_regnum); |
| |
| /* Analyze the instruction at PC. */ |
| nnpc = sparc_analyze_control_transfer (regcache, pc, &npc); |
| if (npc != 0) |
| next_pcs.push_back (npc); |
| |
| if (nnpc != 0) |
| next_pcs.push_back (nnpc); |
| |
| /* Assert that we have set at least one breakpoint, and that |
| they're not set at the same spot - unless we're going |
| from here straight to NULL, i.e. a call or jump to 0. */ |
| gdb_assert (npc != 0 || nnpc != 0 || orig_npc == 0); |
| gdb_assert (nnpc != npc || orig_npc == 0); |
| |
| return next_pcs; |
| } |
| |
| static void |
| sparc_write_pc (struct regcache *regcache, CORE_ADDR pc) |
| { |
| gdbarch *arch = regcache->arch (); |
| sparc_gdbarch_tdep *tdep = gdbarch_tdep<sparc_gdbarch_tdep> (arch); |
| |
| regcache_cooked_write_unsigned (regcache, tdep->pc_regnum, pc); |
| regcache_cooked_write_unsigned (regcache, tdep->npc_regnum, pc + 4); |
| } |
| |
| |
| /* Iterate over core file register note sections. */ |
| |
| static void |
| sparc_iterate_over_regset_sections (struct gdbarch *gdbarch, |
| iterate_over_regset_sections_cb *cb, |
| void *cb_data, |
| const struct regcache *regcache) |
| { |
| sparc_gdbarch_tdep *tdep = gdbarch_tdep<sparc_gdbarch_tdep> (gdbarch); |
| |
| cb (".reg", tdep->sizeof_gregset, tdep->sizeof_gregset, tdep->gregset, NULL, |
| cb_data); |
| cb (".reg2", tdep->sizeof_fpregset, tdep->sizeof_fpregset, tdep->fpregset, |
| NULL, cb_data); |
| } |
| |
| |
| static int |
| validate_tdesc_registers (const struct target_desc *tdesc, |
| struct tdesc_arch_data *tdesc_data, |
| const char *feature_name, |
| const char * const register_names[], |
| unsigned int registers_num, |
| unsigned int reg_start) |
| { |
| int valid_p = 1; |
| const struct tdesc_feature *feature; |
| |
| feature = tdesc_find_feature (tdesc, feature_name); |
| if (feature == NULL) |
| return 0; |
| |
| for (unsigned int i = 0; i < registers_num; i++) |
| valid_p &= tdesc_numbered_register (feature, tdesc_data, |
| reg_start + i, |
| register_names[i]); |
| |
| return valid_p; |
| } |
| |
| static struct gdbarch * |
| sparc32_gdbarch_init (struct gdbarch_info info, struct gdbarch_list *arches) |
| { |
| const struct target_desc *tdesc = info.target_desc; |
| int valid_p = 1; |
| |
| /* If there is already a candidate, use it. */ |
| arches = gdbarch_list_lookup_by_info (arches, &info); |
| if (arches != NULL) |
| return arches->gdbarch; |
| |
| /* Allocate space for the new architecture. */ |
| gdbarch *gdbarch |
| = gdbarch_alloc (&info, gdbarch_tdep_up (new sparc_gdbarch_tdep)); |
| sparc_gdbarch_tdep *tdep = gdbarch_tdep<sparc_gdbarch_tdep> (gdbarch); |
| |
| tdep->pc_regnum = SPARC32_PC_REGNUM; |
| tdep->npc_regnum = SPARC32_NPC_REGNUM; |
| tdep->step_trap = sparc_step_trap; |
| tdep->fpu_register_names = sparc32_fpu_register_names; |
| tdep->fpu_registers_num = ARRAY_SIZE (sparc32_fpu_register_names); |
| tdep->cp0_register_names = sparc32_cp0_register_names; |
| tdep->cp0_registers_num = ARRAY_SIZE (sparc32_cp0_register_names); |
| |
| set_gdbarch_long_double_bit (gdbarch, 128); |
| set_gdbarch_long_double_format (gdbarch, floatformats_ieee_quad); |
| |
| set_gdbarch_wchar_bit (gdbarch, 16); |
| set_gdbarch_wchar_signed (gdbarch, 1); |
| |
| set_gdbarch_num_regs (gdbarch, SPARC32_NUM_REGS); |
| set_gdbarch_register_name (gdbarch, sparc32_register_name); |
| set_gdbarch_register_type (gdbarch, sparc32_register_type); |
| set_gdbarch_num_pseudo_regs (gdbarch, SPARC32_NUM_PSEUDO_REGS); |
| set_tdesc_pseudo_register_name (gdbarch, sparc32_pseudo_register_name); |
| set_tdesc_pseudo_register_type (gdbarch, sparc32_pseudo_register_type); |
| set_gdbarch_pseudo_register_read (gdbarch, sparc32_pseudo_register_read); |
| set_gdbarch_deprecated_pseudo_register_write (gdbarch, |
| sparc32_pseudo_register_write); |
| |
| /* Register numbers of various important registers. */ |
| set_gdbarch_sp_regnum (gdbarch, SPARC_SP_REGNUM); /* %sp */ |
| set_gdbarch_pc_regnum (gdbarch, SPARC32_PC_REGNUM); /* %pc */ |
| set_gdbarch_fp0_regnum (gdbarch, SPARC_F0_REGNUM); /* %f0 */ |
| |
| /* Call dummy code. */ |
| set_gdbarch_frame_align (gdbarch, sparc32_frame_align); |
| set_gdbarch_call_dummy_location (gdbarch, ON_STACK); |
| set_gdbarch_push_dummy_code (gdbarch, sparc32_push_dummy_code); |
| set_gdbarch_push_dummy_call (gdbarch, sparc32_push_dummy_call); |
| |
| set_gdbarch_return_value_as_value (gdbarch, sparc32_return_value); |
| set_gdbarch_stabs_argument_has_addr |
| (gdbarch, sparc32_stabs_argument_has_addr); |
| |
| set_gdbarch_skip_prologue (gdbarch, sparc32_skip_prologue); |
| |
| /* Stack grows downward. */ |
| set_gdbarch_inner_than (gdbarch, core_addr_lessthan); |
| |
| set_gdbarch_breakpoint_kind_from_pc (gdbarch, |
| sparc_breakpoint::kind_from_pc); |
| set_gdbarch_sw_breakpoint_from_kind (gdbarch, |
| sparc_breakpoint::bp_from_kind); |
| |
| set_gdbarch_frame_args_skip (gdbarch, 8); |
| |
| set_gdbarch_software_single_step (gdbarch, sparc_software_single_step); |
| set_gdbarch_write_pc (gdbarch, sparc_write_pc); |
| |
| set_gdbarch_dummy_id (gdbarch, sparc_dummy_id); |
| |
| frame_base_set_default (gdbarch, &sparc32_frame_base); |
| |
| /* Hook in the DWARF CFI frame unwinder. */ |
| dwarf2_frame_set_init_reg (gdbarch, sparc32_dwarf2_frame_init_reg); |
| /* Register DWARF vendor CFI handler. */ |
| set_gdbarch_execute_dwarf_cfa_vendor_op (gdbarch, |
| sparc_execute_dwarf_cfa_vendor_op); |
| /* FIXME: kettenis/20050423: Don't enable the unwinder until the |
| StackGhost issues have been resolved. */ |
| |
| /* Hook in ABI-specific overrides, if they have been registered. */ |
| gdbarch_init_osabi (info, gdbarch); |
| |
| frame_unwind_append_unwinder (gdbarch, &sparc32_frame_unwind); |
| |
| if (tdesc_has_registers (tdesc)) |
| { |
| tdesc_arch_data_up tdesc_data = tdesc_data_alloc (); |
| |
| /* Validate that the descriptor provides the mandatory registers |
| and allocate their numbers. */ |
| valid_p &= validate_tdesc_registers (tdesc, tdesc_data.get (), |
| "org.gnu.gdb.sparc.cpu", |
| sparc_core_register_names, |
| ARRAY_SIZE (sparc_core_register_names), |
| SPARC_G0_REGNUM); |
| valid_p &= validate_tdesc_registers (tdesc, tdesc_data.get (), |
| "org.gnu.gdb.sparc.fpu", |
| tdep->fpu_register_names, |
| tdep->fpu_registers_num, |
| SPARC_F0_REGNUM); |
| valid_p &= validate_tdesc_registers (tdesc, tdesc_data.get (), |
| "org.gnu.gdb.sparc.cp0", |
| tdep->cp0_register_names, |
| tdep->cp0_registers_num, |
| SPARC_F0_REGNUM |
| + tdep->fpu_registers_num); |
| if (!valid_p) |
| return NULL; |
| |
| /* Target description may have changed. */ |
| info.tdesc_data = tdesc_data.get (); |
| tdesc_use_registers (gdbarch, tdesc, std::move (tdesc_data)); |
| } |
| |
| /* If we have register sets, enable the generic core file support. */ |
| if (tdep->gregset) |
| set_gdbarch_iterate_over_regset_sections |
| (gdbarch, sparc_iterate_over_regset_sections); |
| |
| register_sparc_ravenscar_ops (gdbarch); |
| |
| return gdbarch; |
| } |
| |
| /* Helper functions for dealing with register windows. */ |
| |
| void |
| sparc_supply_rwindow (struct regcache *regcache, CORE_ADDR sp, int regnum) |
| { |
| struct gdbarch *gdbarch = regcache->arch (); |
| enum bfd_endian byte_order = gdbarch_byte_order (gdbarch); |
| int offset = 0; |
| gdb_byte buf[8]; |
| int i; |
| |
| /* This function calls functions that depend on the global current thread. */ |
| gdb_assert (regcache->ptid () == inferior_ptid); |
| |
| if (sp & 1) |
| { |
| /* Registers are 64-bit. */ |
| sp += BIAS; |
| |
| for (i = SPARC_L0_REGNUM; i <= SPARC_I7_REGNUM; i++) |
| { |
| if (regnum == i || regnum == -1) |
| { |
| target_read_memory (sp + ((i - SPARC_L0_REGNUM) * 8), buf, 8); |
| |
| /* Handle StackGhost. */ |
| if (i == SPARC_I7_REGNUM) |
| { |
| ULONGEST wcookie = sparc_fetch_wcookie (gdbarch); |
| ULONGEST i7; |
| |
| i7 = extract_unsigned_integer (buf + offset, 8, byte_order); |
| store_unsigned_integer (buf + offset, 8, byte_order, |
| i7 ^ wcookie); |
| } |
| |
| regcache->raw_supply (i, buf); |
| } |
| } |
| } |
| else |
| { |
| /* Registers are 32-bit. Toss any sign-extension of the stack |
| pointer. */ |
| sp &= 0xffffffffUL; |
| |
| /* Clear out the top half of the temporary buffer, and put the |
| register value in the bottom half if we're in 64-bit mode. */ |
| if (gdbarch_ptr_bit (regcache->arch ()) == 64) |
| { |
| memset (buf, 0, 4); |
| offset = 4; |
| } |
| |
| for (i = SPARC_L0_REGNUM; i <= SPARC_I7_REGNUM; i++) |
| { |
| if (regnum == i || regnum == -1) |
| { |
| target_read_memory (sp + ((i - SPARC_L0_REGNUM) * 4), |
| buf + offset, 4); |
| |
| /* Handle StackGhost. */ |
| if (i == SPARC_I7_REGNUM) |
| { |
| ULONGEST wcookie = sparc_fetch_wcookie (gdbarch); |
| ULONGEST i7; |
| |
| i7 = extract_unsigned_integer (buf + offset, 4, byte_order); |
| store_unsigned_integer (buf + offset, 4, byte_order, |
| i7 ^ wcookie); |
| } |
| |
| regcache->raw_supply (i, buf); |
| } |
| } |
| } |
| } |
| |
| void |
| sparc_collect_rwindow (const struct regcache *regcache, |
| CORE_ADDR sp, int regnum) |
| { |
| struct gdbarch *gdbarch = regcache->arch (); |
| enum bfd_endian byte_order = gdbarch_byte_order (gdbarch); |
| int offset = 0; |
| gdb_byte buf[8]; |
| int i; |
| |
| /* This function calls functions that depend on the global current thread. */ |
| gdb_assert (regcache->ptid () == inferior_ptid); |
| |
| if (sp & 1) |
| { |
| /* Registers are 64-bit. */ |
| sp += BIAS; |
| |
| for (i = SPARC_L0_REGNUM; i <= SPARC_I7_REGNUM; i++) |
| { |
| if (regnum == -1 || regnum == SPARC_SP_REGNUM || regnum == i) |
| { |
| regcache->raw_collect (i, buf); |
| |
| /* Handle StackGhost. */ |
| if (i == SPARC_I7_REGNUM) |
| { |
| ULONGEST wcookie = sparc_fetch_wcookie (gdbarch); |
| ULONGEST i7; |
| |
| i7 = extract_unsigned_integer (buf + offset, 8, byte_order); |
| store_unsigned_integer (buf, 8, byte_order, i7 ^ wcookie); |
| } |
| |
| target_write_memory (sp + ((i - SPARC_L0_REGNUM) * 8), buf, 8); |
| } |
| } |
| } |
| else |
| { |
| /* Registers are 32-bit. Toss any sign-extension of the stack |
| pointer. */ |
| sp &= 0xffffffffUL; |
| |
| /* Only use the bottom half if we're in 64-bit mode. */ |
| if (gdbarch_ptr_bit (regcache->arch ()) == 64) |
| offset = 4; |
| |
| for (i = SPARC_L0_REGNUM; i <= SPARC_I7_REGNUM; i++) |
| { |
| if (regnum == -1 || regnum == SPARC_SP_REGNUM || regnum == i) |
| { |
| regcache->raw_collect (i, buf); |
| |
| /* Handle StackGhost. */ |
| if (i == SPARC_I7_REGNUM) |
| { |
| ULONGEST wcookie = sparc_fetch_wcookie (gdbarch); |
| ULONGEST i7; |
| |
| i7 = extract_unsigned_integer (buf + offset, 4, byte_order); |
| store_unsigned_integer (buf + offset, 4, byte_order, |
| i7 ^ wcookie); |
| } |
| |
| target_write_memory (sp + ((i - SPARC_L0_REGNUM) * 4), |
| buf + offset, 4); |
| } |
| } |
| } |
| } |
| |
| /* Helper functions for dealing with register sets. */ |
| |
| void |
| sparc32_supply_gregset (const struct sparc_gregmap *gregmap, |
| struct regcache *regcache, |
| int regnum, const void *gregs) |
| { |
| const gdb_byte *regs = (const gdb_byte *) gregs; |
| gdb_byte zero[4] = { 0 }; |
| int i; |
| |
| if (regnum == SPARC32_PSR_REGNUM || regnum == -1) |
| regcache->raw_supply (SPARC32_PSR_REGNUM, regs + gregmap->r_psr_offset); |
| |
| if (regnum == SPARC32_PC_REGNUM || regnum == -1) |
| regcache->raw_supply (SPARC32_PC_REGNUM, regs + gregmap->r_pc_offset); |
| |
| if (regnum == SPARC32_NPC_REGNUM || regnum == -1) |
| regcache->raw_supply (SPARC32_NPC_REGNUM, regs + gregmap->r_npc_offset); |
| |
| if (regnum == SPARC32_Y_REGNUM || regnum == -1) |
| regcache->raw_supply (SPARC32_Y_REGNUM, regs + gregmap->r_y_offset); |
| |
| if (regnum == SPARC_G0_REGNUM || regnum == -1) |
| regcache->raw_supply (SPARC_G0_REGNUM, &zero); |
| |
| if ((regnum >= SPARC_G1_REGNUM && regnum <= SPARC_O7_REGNUM) || regnum == -1) |
| { |
| int offset = gregmap->r_g1_offset; |
| |
| for (i = SPARC_G1_REGNUM; i <= SPARC_O7_REGNUM; i++) |
| { |
| if (regnum == i || regnum == -1) |
| regcache->raw_supply (i, regs + offset); |
| offset += 4; |
| } |
| } |
| |
| if ((regnum >= SPARC_L0_REGNUM && regnum <= SPARC_I7_REGNUM) || regnum == -1) |
| { |
| /* Not all of the register set variants include Locals and |
| Inputs. For those that don't, we read them off the stack. */ |
| if (gregmap->r_l0_offset == -1) |
| { |
| ULONGEST sp; |
| |
| regcache_cooked_read_unsigned (regcache, SPARC_SP_REGNUM, &sp); |
| sparc_supply_rwindow (regcache, sp, regnum); |
| } |
| else |
| { |
| int offset = gregmap->r_l0_offset; |
| |
| for (i = SPARC_L0_REGNUM; i <= SPARC_I7_REGNUM; i++) |
| { |
| if (regnum == i || regnum == -1) |
| regcache->raw_supply (i, regs + offset); |
| offset += 4; |
| } |
| } |
| } |
| } |
| |
| void |
| sparc32_collect_gregset (const struct sparc_gregmap *gregmap, |
| const struct regcache *regcache, |
| int regnum, void *gregs) |
| { |
| gdb_byte *regs = (gdb_byte *) gregs; |
| int i; |
| |
| if (regnum == SPARC32_PSR_REGNUM || regnum == -1) |
| regcache->raw_collect (SPARC32_PSR_REGNUM, regs + gregmap->r_psr_offset); |
| |
| if (regnum == SPARC32_PC_REGNUM || regnum == -1) |
| regcache->raw_collect (SPARC32_PC_REGNUM, regs + gregmap->r_pc_offset); |
| |
| if (regnum == SPARC32_NPC_REGNUM || regnum == -1) |
| regcache->raw_collect (SPARC32_NPC_REGNUM, regs + gregmap->r_npc_offset); |
| |
| if (regnum == SPARC32_Y_REGNUM || regnum == -1) |
| regcache->raw_collect (SPARC32_Y_REGNUM, regs + gregmap->r_y_offset); |
| |
| if ((regnum >= SPARC_G1_REGNUM && regnum <= SPARC_O7_REGNUM) || regnum == -1) |
| { |
| int offset = gregmap->r_g1_offset; |
| |
| /* %g0 is always zero. */ |
| for (i = SPARC_G1_REGNUM; i <= SPARC_O7_REGNUM; i++) |
| { |
| if (regnum == i || regnum == -1) |
| regcache->raw_collect (i, regs + offset); |
| offset += 4; |
| } |
| } |
| |
| if ((regnum >= SPARC_L0_REGNUM && regnum <= SPARC_I7_REGNUM) || regnum == -1) |
| { |
| /* Not all of the register set variants include Locals and |
| Inputs. For those that don't, we read them off the stack. */ |
| if (gregmap->r_l0_offset != -1) |
| { |
| int offset = gregmap->r_l0_offset; |
| |
| for (i = SPARC_L0_REGNUM; i <= SPARC_I7_REGNUM; i++) |
| { |
| if (regnum == i || regnum == -1) |
| regcache->raw_collect (i, regs + offset); |
| offset += 4; |
| } |
| } |
| } |
| } |
| |
| void |
| sparc32_supply_fpregset (const struct sparc_fpregmap *fpregmap, |
| struct regcache *regcache, |
| int regnum, const void *fpregs) |
| { |
| const gdb_byte *regs = (const gdb_byte *) fpregs; |
| int i; |
| |
| for (i = 0; i < 32; i++) |
| { |
| if (regnum == (SPARC_F0_REGNUM + i) || regnum == -1) |
| regcache->raw_supply (SPARC_F0_REGNUM + i, |
| regs + fpregmap->r_f0_offset + (i * 4)); |
| } |
| |
| if (regnum == SPARC32_FSR_REGNUM || regnum == -1) |
| regcache->raw_supply (SPARC32_FSR_REGNUM, regs + fpregmap->r_fsr_offset); |
| } |
| |
| void |
| sparc32_collect_fpregset (const struct sparc_fpregmap *fpregmap, |
| const struct regcache *regcache, |
| int regnum, void *fpregs) |
| { |
| gdb_byte *regs = (gdb_byte *) fpregs; |
| int i; |
| |
| for (i = 0; i < 32; i++) |
| { |
| if (regnum == (SPARC_F0_REGNUM + i) || regnum == -1) |
| regcache->raw_collect (SPARC_F0_REGNUM + i, |
| regs + fpregmap->r_f0_offset + (i * 4)); |
| } |
| |
| if (regnum == SPARC32_FSR_REGNUM || regnum == -1) |
| regcache->raw_collect (SPARC32_FSR_REGNUM, |
| regs + fpregmap->r_fsr_offset); |
| } |
| |
| |
| /* SunOS 4. */ |
| |
| /* From <machine/reg.h>. */ |
| const struct sparc_gregmap sparc32_sunos4_gregmap = |
| { |
| 0 * 4, /* %psr */ |
| 1 * 4, /* %pc */ |
| 2 * 4, /* %npc */ |
| 3 * 4, /* %y */ |
| -1, /* %wim */ |
| -1, /* %tbr */ |
| 4 * 4, /* %g1 */ |
| -1 /* %l0 */ |
| }; |
| |
| const struct sparc_fpregmap sparc32_sunos4_fpregmap = |
| { |
| 0 * 4, /* %f0 */ |
| 33 * 4, /* %fsr */ |
| }; |
| |
| const struct sparc_fpregmap sparc32_bsd_fpregmap = |
| { |
| 0 * 4, /* %f0 */ |
| 32 * 4, /* %fsr */ |
| }; |
| |
| void _initialize_sparc_tdep (); |
| void |
| _initialize_sparc_tdep () |
| { |
| gdbarch_register (bfd_arch_sparc, sparc32_gdbarch_init); |
| } |